Low pH and Anionic Lipid-dependent Fusion of Uukuniemi Phlebovirus to Liposomes

Abstract

Many phleboviruses (family Bunyaviridae) are emerging as medically important viruses. These viruses enter target cells by endocytosis and low pH-dependent membrane fusion in late endosomes. However, the necessary and sufficient factors for fusion have not been fully characterized. We have studied the minimal fusion requirements of a prototypic phlebovirus, Uukuniemi virus, in an in vitro virus-liposome assay. We show that efficient lipid mixing between viral and liposome membranes requires close to physiological temperatures and phospholipids with negatively charged headgroups, such as the late endosomal phospholipid bis(monoacylglycero)phosphate. We further demonstrate that bis(monoacylglycero)phosphate increases Uukuniemi virus fusion beyond the lipid mixing stage. By using electron cryotomography of viral particles in the presence or absence of liposomes, we observed that the conformation of phlebovirus glycoprotein capsomers changes from the native conformation toward a more elongated conformation at a fusion permissive pH. Our results suggest a rationale for phlebovirus entry in late endosomes.

Keywords: bis(monoacylglycero)phosphate; bunyavirus; electron tomography; membrane; membrane fusion; phlebovirus; virus entry; virus structure.

FIGURE 1.

Electron cryotomography and subtomogram averaging of UUKV virus. A, a two-dimensional projection image of Uukuniemi (UUKV) virions at pH 7.5 collected at −6 μm defocus at 0° tilt. Gold nanoparticles (∼10 nm in diameter) have been added as fiducial markers to align images in tomographic tilt series. Virions were ∼110 nm in diameter and covered in glycoprotein capsomers. Smaller particles (∼50 nm) with virion-like morphology were occasionally observed (arrowhead). B, a 10-nm thick slice through a three-dimensional tomographic reconstruction of UUKV virions at pH 7.5 collected at −6 μm defocus. Scale bar = 100 nm for A and B. C, a surface rendering of the viral glycoprotein capsomers is shown for UUKV (left) and RVFV (right, EMDB-1550) solved using subtomogram averaging (this study, UUKV) and single particle icosahedral averaging (RVFV) (18). In both cases, one hexameric capsomer that is surrounded by six neighboring capsomers is shown. D and E, slices (6 nm thick) through the density maps are shown parallel (D) and orthogonal (E) to the viral membrane (M). Dimensions of the capsomers are indicated. Scale bar = 15 nm for C–E.

FIGURE 2.

Membrane labeling of UUKV and SFV with pyrene. A and B, fluorescence emission spectra (excitation at 345 nm) are shown for pyrene-labeled Uukuniemi virus (UUKV-pyr) and Semliki Forest virus (SFV-pyr). Spectra were measured before (black) and after (gray) solubilizing the membrane with 0.2% (w/v) Triton X-100 (TX-100). Addition of Triton X-100 results in the solubilization of the membrane, and thus further dissociation of dimeric pyrene excimers. This is seen as increased emission (up arrows) at 378 and 395 nm corresponding to the pyrene monomers and as disappearance (down arrows) of the broad emission peak at 475 nm corresponding to the pyrene dimers.

FIGURE 3.

BMP promotes lipid mixing between UUKV and liposomes. A and B, pyrene-labeled Uukuniemi virus (UUKV-pyr) was mixed with either DOPC-DOPE (Lipo w/o BMP) or DOPC-DOPE-BMP (Lipo w BMP) liposomes and preincubated at 37 °C. After acidification to pH 5, lipid mixing was followed over time. In the presence of BMP, efficient lipid mixing was observed and, after 600 s, a plateau was reached at lipid mixing index of ∼0.6. No significant lipid mixing was observed in the absence of BMP. The first 25 s (indicated with a box in A) are shown in B. Most of the lipid mixing occurred in the first few seconds after acidification. C, quantification of UUKV-pyr and pyrene-labeled Semliki Forest virus (SFV-pyr; control) lipid mixing at 5 min post-acidification in the presence of liposomes. In addition to control DOPC-DOPE liposomes (Lipo), liposomes with added lipids were tested as indicated. As expected, SFV lipid mixing required both sphingomyelin (SPM) and cholesterol (Chol), whereas UUKV lipid mixing was only observed in the presence of BMP. D, quantification of SFV-pyr lipid mixing after 5 min at pH 5.5. No significant lipid mixing was observed in the presence of BMP and absence of SPM and Chol. Note that the y axis is on a different scale in C and D. Error bars represent 1 S.D. (n = 3). Statistical significance was determined using a Student's t test using significance levels: *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001.

FIGURE 4.

BMP promotes UUKV-induced fusion of BHK-21 cells. A, UUKV-induced cell-cell fusion visualized by fluorescence confocal microscopy. UUKV was added to BHK-21 cells, cell-cell fusion was induced by acidification to pH 5.0, and cell nuclei (blue) and plasma membrane (orange) were stained for counting the number of syncytia (i.e. cells with multiple nuclei). As a negative control, cells were first imaged without added virus or BMP (1). In the absence of added BMP, and in the presence of UUKV, some syncytia could be observed (2). The effect of BMP on syncytia formation was tested by adding BMP either before (3) or after (4) acidification. Scale bar = 25 μm. B, quantification of cell-cell fusion. Labels 1–4 are as in A. Note that BMP significantly increased syncytia formation only when added before acidification. Error bars represent the mean ± S.E. of the fusion index. Statistical significance was determined using a Student's t test using significance levels: *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001.

FIGURE 5.

UUKV glycoproteins are required for lipid mixing. A, fluorescence emission spectrum of liposomes (Lipo-pyr) with a phospholipid composition similar to pyrene-labeled UUKV with and without added Triton X-100. B, quantification of lipid mixing between pyrene-labeled liposomes (Lipo-pyr) or Lipo-pyr with 2.5% (w/v) sucrose and unlabeled liposomes with BMP at 10 min post-acidification, pH 5.0. Pyrene-labeled Uukuniemi virus (UUKV-pyr) was added as a positive control for lipid mixing. No significant lipid mixing was observed in the case of labeled donor liposomes. Error bars represent 1 S.D. (n = 3). Statistical significance was determined using a Student's t test using significance levels: *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001.

FIGURE 6.

UUKV lipid mixing with liposomes is pH and temperature sensitive. A, lipid mixing of pyrene-labeled Uukuniemi virus (UUKV-pyr) with BMP containing liposomes was determined between pH 4.0 and 7.5 at 37 °C. For accurate pH measurements, pH was measured after acidification for each data point. Error bars represent 1 S.D. for both the pH (horizontal bars) and lipid mixing index (vertical bars; n = 3). A sigmoidal curve was fitted and the confidence of the fit is indicated with dotted lines (95% confidence level). Half-maximal lipid mixing was achieved at pH 5.6 (dashed lines). B, lipid mixing of UUKV-pyr with BMP-containing liposomes was determined at pH 5.0 at different temperatures. Error bars represent 1 S.D. (n = 4). Statistical significance was determined using a Student's t test using significance levels: *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001.

FIGURE 7.

UUKV induces content release from liposomes in the presence of BMP. A and B, lipid mixing of pyrene-labeled Uukuniemi virus (UUKV-pyr) with DOPC-DOPE liposomes labeled with SRB (Lipo-SRB) was determined at pH 5.0, 37 °C. In the presence of BMP, a rapid UUKV-induced content release was observed as shown in A. In the absence of BMP, content release remained at background levels as shown in B. Notice that also in the absence of UUKV (black trace) some spontaneous release of SRB was observed. C, quantification of content release at 600 s post-acidification confirmed that UUKV-induced content release was significantly higher in the presence of BMP than in its absence. Furthermore, UUKV-induced content release was significantly higher than the spontaneous release of SRB dye from the liposomes. Error bars represent 1 S.D. (n = 3 for Lipo-SRB and n = 4 for Lipo-SRB with BMP). Statistical significance was determined using a Student's t test using significance levels: *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001. D, simultaneous measurement of content release (black trace) and lipid mixing (gray trace) in the presence of UUKV-pyr and BMP containing liposomes labeled with SRB showed that content release starts concomitantly to lipid mixing and follows roughly the same kinetics.

FIGURE 8.

UUKV lipid mixing with liposomes containing non-physiological anionic lipids. A, pyrene-labeled Uukuniemi virus (UUKV-pyr) lipid mixing was measured with control DOPC-DOPE liposomes (Lipo) and with three other DOPC-DOPE liposomes containing additional phospholipids with negatively charged head groups, namely PG, PS, and PA. After acidification to pH 5.0, lipid mixing was followed over time. In the presence of PA and PG, efficient lipid mixing was observed. In the presence of PS, lipid mixing was at background level. B, quantification of UUKV-pyr lipid mixing at 5 min post-acidification in the presence of the same liposomes as in A. Liposomes with PG and PA showed significantly higher lipid mixing than the control liposomes (Lipo). Error bars represent 1 S.D. (n = 3). Statistical significance was determined using a Student's t test using significance levels: *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001.

FIGURE 9.

Electron cryotomography of UUKV fusion. A and B, shown are 10-nm thick tomographic slices of UUKV (V) with liposomes (L) at either pH 7.5 (control) or 5.0. Liposomes shown contained BMP and liposomes in the lower panel in B had also added cholesterol (see Table 1). Tomograms have been low-pass filtered to 6 nm⁻¹ spatial frequency to emphasize the membranes and capsomers. Scale bar = 50 nm. C, a close-up of the areas indicated in B are shown. Regions with extended capsomer structures (magenta) connecting the virus membrane (yellow) to the liposome membrane (cyan) are indicated with arrowheads in the panel on the left and colored in the panel on the right. For size comparison, five G_C monomers in an extended conformation (Protein Data Bank code 4HJC) (14) were placed side by side in different orientations around the long axis and filtered to the same spatial frequency (inset). Note that such extended structures were also observed in the absence of BMP. Scale bar = 15 nm. D, a dimple structure at the virus-liposome contact site (upper panel) and a putative full fusion site where viral and liposome membranes have fully merged (bottom panel) are indicated with arrowheads. Scale bar = 50 nm.